fix: update @asyncapi/optimizer to 0.3.0 version

[asyncapi-bot]

No description provided.

[sonarqubecloud]

Quality Gate passed

Kudos, no new issues were introduced!

0 New issues
0 Security Hotspots
No data about Coverage
0.0% Duplication on New Code

See analysis details on SonarCloud

[derberg]

@KhudaDad414 can you have a look, in theory tests should pass as there were no breaking changes

[KhudaDad414]

@derberg, just updating the tests over here. Quick thing though, I'm a bit puzzled about how we should handle versioning for our optimizer. Since we've switched to parser-js, what we can do now depends on the components we have at hand. For instance, we can't parse parameters from components anymore, which means we aren't optimizing parameters like before. The optimizer's interface is the same, sure, but the output it's spitting out is different now.

I'm thinking, are the tests where we expect the same output for the same input still the right way to go? Feels like we might need a testing approach that's a bit more adaptable, not just expecting the same results every time. Also, there's this thing about whether we need to bump up to a major version if we start supporting a new type of component since the output changes? What do you think?

[derberg]

@KhudaDad414 actually the change of output is also a breaking change.

optimizer, bundler, modelina -> breaking change is not only library API change but they output as well as in case of these, the output is what user depends on and expects to not have changes that can break things. It is like with any API like REST API, you do not consider just resource name change a breaking change, but if API response changes, it is breaking change as well.

so if output changes in the way it breaks things, it is a breaking change

look into #573

Since we've switched to parser-js, what we can do now depends on the components we have at hand. For instance, we can't parse parameters from components anymore, which means we aren't optimizing parameters like before

then maybe parser needs improvement. Although it is different topic I think and here just used as an example